The current UK election is interesting in that unlike 2017, this time it is not just a two horse race (the size of the Liberal Democrat support this time round – despite decline and rumours of collapse – adds unpredictability to the seats per party). Also the Brexit/Remain vote cuts right through the traditional left/right vote, and both are near as dammit 50% so you are really watching 4 “virtual parties” – Brexit Right, Brexit Left, Remain Right and Remain Left – duke it out.
Anyway, we have been tracking it at the estimated “% support per party” level using our DataSwarm system, which essentially tracks what we call the Zeitgeist – the spirit of the time – as evidenced in the volume, velocity and sentiment of people’s conversation. This has proven very predictive over a number of elections we have correctly forecast, often against the run of the polls (see more on this over here).
At any rate, we have translated this into a “% share” snapshot today, as with less than a week to go, the end state usually starts to emerge and this is shown in the chart at the top. (There are the standard Weaselly Caveats to this prediction – see the bottom of the page). The chart only shows the 4 major parties, who together right now compromise c 95% of the share.
As of today (7th Dec ) it shows the Tory lead slowly growing and the system is predicting they will win the election, but not with an overall majority. But (and this is a big “but”, see the Weaselly Caveats below) the Tories have a slow rise in their dominance and this momentum can reinforce itself – the “big Mo” as the US calls it – and means they may get over the 50% of seats line. The trend has increased after last night’s BBC Debate, but this may well be an event-caused blip. Alternatively, given the margins are so tight, a relatively small shift means a potential Lib/Lab coalition being big enough to rule is an almost equal possibility.
For what its worth most of the system’s tracking algorithms still predict the Tories will undershoot the 50% mark, but will come in with enough of a difference to be larger than the next 2 parties combined.
So overall, we thus have a high probability of a hung Parliament, but it is less clear who is going to be hanging….
Which of course makes the probability of “Getting Brexit Done” very unclear and we may be right back here in the near future.
Anyway, interesting days ahead…..
- Predictions about the future should always have margins of error attached, Tories and Labour are still (just) within 0.95 deviation levels of position reversal (ie there is a c 5% chance they may reverse position.)
- The likelihood of the Tories increasing their position to an outright win is still relatively high, but it is still “very possible” (within 1 standard deviation of trend growth), not “probable”.
- About as likely is their gap drops, and Liberals + Labour end up slightly more votes
- In 2017 the system dealt with the “First Past The Post”, “Tactical Voting” and “Shy Tory” distortions fairly well but these error drivers are far larger this time:
- Liberal Democrats (potentially) have a far larger impact.
- Multi Axis Tactical Voting – Brexit/Remain, Left/Right etc.
- It’s not clear just what a Tory is these days, never mind a shy one….